Will The 2016 Blue Jays Rotation Be Better Than 2015?

There are a multitude of ways to analyze any given aspect of an MLB team in order to get insight into what might be expected from the coming year. Despite that, no one really knows anything for sure¹, it’s all just a matter of probabilities. Over the last decade, those probabilities have only become more complex as we’ve been drowned in information provided by advanced metrics and complex analytics. Of course, no one ever accused me of being advanced and complex when it comes to the math.

That, however, does not keep me from attempting to apply my own brand of creativity to the question of year-to-year team analysis. What follows is a perhaps unorthodox but, I believe, informative thesis for comparing the potential of the Blue Jays’ 2016 starting rotation to the production that the team received in 2015. Underlying this comparison is the assumption that…

1. The offense stands to be at least as good as it was last season
2. The defense projects to be considerably better.

The thesis I propose is this – that the five turns in a standard rotation may be imagined to be five boxes, or baskets, which each contain the production of that individual turn over the course of a season. For example, R.A. Dickey took every turn for his spot in the rotation in 2015, thus one of the baskets for 2015 is filled exclusively with the work of RA Dickey. Likewise, since he is still here and counted on to produce, one may assume that one of the baskets for 2016 can be projected based on what we can reasonably expect Dickey to put into it. So my plan is to first see what was in each basket in 2015, and then consider whether it is reasonable to expect the candidates currently on-hand to replicate (or exceed) those contents in the upcoming season.

The rotation that opened the season last year was as follows: Drew Hutchison, R.A. Dickey, Daniel Norris, Mark Buehrle, Aaron Sanchez. Now while it is obviously not the case that those appeared in order of best through worst (as might have been presumed last March) and therefore not 1, 2, etc in order, it is nevertheless the order in which the baskets were respectively filled as the season played out. So Hutchinson begins to fill the first and so forth .Over the course of the season, two baskets were filled by only one pitcher, two were filled by two pitchers each, and the fifth was a grab bag of six others. There are two technical variants from this pattern.

On May 19, the team decided to insert one Todd Redmond start to give the rotation an extra day’s rest (an event that perhaps coincidentally served as a big statistical turning point for the whole staff). I choose to include this in the fifth basket. Similarly, on August 2 the team opted to start R.A. Dickey on 3 days’ rest in order to line up David Price’s starts as they wanted them. This went in Dickey’s basket and was ignored in terms of rotation days otherwise.

So, what did the five baskets contain in 2015?

As I mentioned, the first is primarily filled with the maddening contributions of Drew Hutchison. But it also contains Marcus Stroman’s four superlative starts. The results look like this.

The second, of course, is all Dickey.

The third, began with Daniel Norris, but contains both his work and that of Marco Estrada.

The fourth, nothing but Buehrle.

And finally, the fifth is filled by Sanchez, and then a succession of pretenders until David Price was acquired.

Now, in order to make the best comparison I can between these, and what they might look like in 2016, I will reorder them in rough order of their effectiveness (the middle three could be in almost any order).

Now we come to the question: can that level of success be repeated with the current personnel? To answer this, I will compare each 2015 basket with two perspectives on what one might expect next year, and the stat line projected by the Steamer system found on Fangraphs.

Basket 1

Expected, barring another freak occurrence, to be Marcus Stroman. What follows then, in the first table, is the line from last year, compared with the stats from Stroman’s career as a SP to date, then a line which pro-rates those totals to 33 starts, and finally those projected by Steamer (note that counting stats other than starts and IP are not shown in those projections)

Both the objective analysis, and my own “back-of-the-napkin” guess that Stroman should at least replicate his previous rate of production agree that it is entirely reasonable to expect Stroman to equal, if not surpass, the 2015 production of Basket 1. I quibble with the machine’s ERA projection in that the other rate stats don’t seem to support the inflation but that’s beyond my pay grade. Also note that the system projects Stroman for 3.5 WAR according to this production.

Basket 2

This one was Dickey before, and should be Dickey again. The first line is last year, the next is the average production of his three years as a Blue Jays, and the last is Steamer.

Again, there’s really little dispute here. The machine factors in age regression which is typically valid but one can reasonably argue Dickey is an exception to that sort of calculation. Dickey posted 2 WAR last season, only slightly higher than his 2 previous seasons (1.7 each) and while the machine projects 1.1, it is said to be driven by FIP which is a stat Dickey “breaks” every year. I’d pencil him in for something between .5 and .7 more than what the machine says. In other words, yes, he can re-fill his basket just fine.

Basket 3

This one is Marco Estrada’s to fill and here we begin to get into a bit more speculation given that Marco so completely astonished everyone last season. The basis for comparison here is the group effort that made up Basket 5 last year, so the bar is not quite so high. Behold and be amazed. The top line you know, the second line is a 32-start average based on Estrada’s 99 career starts. The last is Steamer, pro-rated to 32 starts (they actually project him at 29).

Look at how well Estrada matches up here: fewer hits, more homers, but otherwise near identical lines between his career pattern, and the expectations established by the comparison line. Steamer drops his K rate for some reason I cannot discern, and like Dickey, he completely busted xFIP last year, and his FIP and xFIP were perfectly fine in 2012 and 2013 so I’ll mildly dispute the regression the machine levies, and seriously dispute the projection of a mere .5 WAR in the upcoming season. His 1.8 last year mirrored the 1.8 in 2013, and he spiked to 3.3 in 2012. I see no reason to project a drop in that for 2016.  So we have a third basket that we can reasonably assume will turn out at least as well as last year.

Basket 4

What was previously filled by Mark Buehrle falls to JA Happ. Is he up to it? I argue he certainly is. Let’s acknowledge up front that there is a difference in innings. Buehrle’s signature accomplishment is his reliable provision of 200 IP every year, and last year he averaged 6.2 IP per start. But we’ll lay that aside for just a moment. I know what you are thinking right now, “She’s about to hang her whole case on Happ’s aberrational Pittsburgh run last year, which would be foolish.” WRONG! Well, it’s not wrong that it would be foolish to rest my whole argument on Pittsburgh, but it is wrong that the second half of 2015 is where I intend to build my case. For the purpose of this exercise, I intend to completely set aside what happened in Pittsburgh except to note it as a positive indicator that the performance I’m suggesting is a reasonable expectation.

Rather, let’s look at JA Happ for the previous full year (and a bit more). Specifically, from the 2014 ASB through (almost) the same date in 2015. This sample includes 32 starts, over a span of time in which his team played 160 games. The comparison you see on the chart below is Buehrle’s production, compared to this 32 start sample of Happ’s work, and the Steamer figures (pro-rated from 28 to 32 starts).

As I alluded to, Buehrle does have the IP factor, but he averaged 6.2 per start in 2015, while in this sample Happ came out at 5.8 per start. Over the course of a season, that’s less than 2 outs per game (Steamer has him averaging 5.5 per start, but over his career in all starts he’s averaged 5.67). That IP/S figure, by the way, mirrors the rate he posted in Pittsburgh.  As you surely know, the rest of his rate stats were much better for the Pirates. I can’t argue the foregoing year should cause you to dream on stats like those in Pittsburgh, but the latter should give you confidence to believe in the former. So the conclusion here is that, apart from slightly fewer innings, Happ pitched for a full year (part of it spent in Toronto) in a fashion very similar in value to the year we got from Buehrle in 2015. Steamer values Happ’s projection at (pro-rated) about 2 WAR which is right in line with the 2.1 Buehrle posted. Four baskets in and we’re still good.

Basket 5

And now we come to the real wild card of the staff – Drew Hutchison. Before I go to the numbers, I’m going to go ahead and stake out a base-line claim: Drew is wildly unlikely to get worse than he was last year. There are good indicators (excellent home results, better xFIP than ERA) that a better pitcher is in there. Most significant among these is that he was over one full run better, as measured by ERA, in 2014. For a sample set for comparison in this case, I simply took the 162 game average on his career rates to this point. A note about the Steamer line here – they only project Hutch for 15 starts which seems completely counter-intuitive given the numbers they came up with otherwise, and his relative durability the last two years. I’ve pro-rated them to 32 starts, as I did with Estrada and Happ.

Since the bulk of the 2015 basket contains Hutchison innings, it stands to reason there would be a lot of similarity here. Across the board, the rates are indeed quite close, but I note with interest that Steamer is actually optimistic about his ERA (presumably, another instance of the impact of FIP on their projections). In any case, while the bar is low here there’s no reason at all to doubt Hutch’s ability to clear it.The pro-rated WAR from the Steamer projection comes out to 1.7, just a bit higher than his 1.5 in 2015 (and likely essentially identical to the Hutch/Stroman combination of stats reflected in the top line of the last chart). Again, that makes me wonder why they cede so many of his starts to Chavez.

Conclusions

There’s your basic analysis – if you got through the year on your Opening Day starting five, it’s essentially the same rotation as last year in terms of projected production. But of course, you never actually do that, right?  The Blue Jays used 10 starters in all, six of them were essentially the primary rotation. Discounting mid-season acquisition David Price, the Jays only used an “alternative” starter 16 times.Over the entire season, all starters combined for 12.5 WAR according to Fangraphs. Collectively, my estimation above ranges somewhere close to 11. That’s about 2.2-2.4 higher than the total WAR for SP in the Fangraph projections.

This is where the comparisons break down just a bit. Obviously the odds are low that the Blue Jays will actually make it through the year with only five starters. FanGraphs projects the nominal 6th starter, Jesse Chavez, for something like 22-24 starts valued at 1.8 WAR². Obviously this contribution would have to come out of the accumulated totals in my projection for the front five. The good news, though, is that such a projection puts him on pace with the work of any of the others, excepting Stroman. So feel free to sub him in for Hutch if you prefer. Either way, there’s a bit of a marginal difference in WAR that I can’t account for and dismiss as mostly a product of me mashing up stat projections in a way they were not intended to be used.

In conclusion and summary, repeating the results of 2015 depends in large measure on Stroman being as good as we think he is based on past performance (if he’s as good as HE thinks he is, he’s a Cy caliber pitcher). Beyond that, it’s possible that Dickey declines a bit, reasonable that Estrada regresses some, but both within margins that don’t really affect these calculations. Happ is legitimate, even if he ends up not being a 3+ WAR pitcher ever again, and Hutchison is the X factor. He shouldn’t be worse and he could very well be better – considerably better.

And I haven’t even mentioned Sanchez yet. Per Marcus Stroman, he’s bulked up with full intention of being a starter. In theory he could take the No. 5 spot and push Hutch back to Buffalo to start the year, which might be a good measure of whether or not Drew has gotten a handle on his issues. As much as I’m a believer in Hutchison as a mid-rotation starter at his peak, it would not kill us or him for that to happen. With this said, I did not include Sanchez as an alternative above because there are just too many unknowns. If both pitch well in the Spring, the Blue Jays will have to decide whether it’s better to strengthen the ‘pen with Sanchez (with the potential consequences others have noted) or strengthen their depth by having an good young experienced starter in reserve in AAA. But the more I speak of this the more I drift off topic. So, I’ll stop now, and conclude that the Jays have a good shot at having similar production in the rotation, with upside for more. In the end, that’s probably good enough.

Steamer Projections via FanGraphs

Feature Image via Wikipedia Commons

The post Will The 2016 Blue Jays Rotation Be Better Than 2015? appeared first on Blue Jays Plus.